SESSION TYPE: ICU Infections
PRESENTED ON: Wednesday, October 24, 2012 at 02:45 PM - 04:15 PM
PURPOSE: An association between increased venous thromboembolic (VTE) events and H1N1 Influenza A (H1N1) was noted in the first 10 patients admitted to the University of Michigan Surgical Intensive Care Unit in the spring of 2009, prompting initiation of an empiric full-dose anticoagulation protocol for H1N1 suspected cases.
METHODS: A cohort of critically ill ARDS patients (n=71) were identified during 2009-2010, of which 36 patients had confirmed H1N1 infection. Measured parameters included: age, APACHE Score, body mass index, anticoagulation relative to VTE diagnosis, PaO2/FiO2 ratio, bacterial pneumonia, ICU length of stay, hospital mortality, complications from anticoagulation, and days on ECMO, ventilator, pressors, and continuous renal replacement therapy. Endpoints were any VTE (DVT, PE, mesenteric thrombosis) and pulmonary thromboembolism. Univariate comparisons and multivariate logistic regression modeling was used to identify risk factors for thrombosis using STATA software (College Park, TX). Thrombotic event-free survival was analyzed using Kaplan Meier analysis and log rank test.
RESULTS: Multivariate logistic regression identified the following independent risk factors for any thrombotic event: H1N1 (OR 17.9 (95% CI 2.4-130.6), p=0.004), culture positive bacterial pneumonia (OR 6 (95% CI 1.6-22.9) p=0.008) and vasopressor requirement (OR 13.1 (95% CI 1.4-119.1), p=0.022). Independent risk factors for pulmonary embolism were: H1N1 (OR 23.3 (95% CI 2.1-260.9) p=0.011), culture positive bacterial pneumonia (OR 10.4 (95% CI 1.7-62.2) p=0.01), and male sex (OR 8.8 (95% CI 1.1-76.2) p=0.048). Empiric anticoagulation conferred protection against both outcomes (p < 0.02). Kaplan meier analysis showed patients with H1N1 infection not empirically anticoagulated were at the highest risk for thrombotic events (p<0.0001). Stratified log rank test confirmed that empiric anticoagulation provided significant protection from thrombotic events only in the H1N1 subset of the ARDS cohort (p<0.0001).
CONCLUSIONS: Critically ill patients with severe ARDS due to H1N1 influenza have increased risk of venous thromboembolic complications, particularly pulmonary thromboembolism. Empiric anticoagulation significantly reduced incidence of thrombosis without increased hemorrhagic complications.
CLINICAL IMPLICATIONS: This evidence strongly supports initiation of full-dose anticoagulation in critically ill ARDS patients with suspected H1N1 Influenza A.
DISCLOSURE: The following authors have nothing to disclose: Benjamin Jacobs, Andrea Obi, Shipra Arya, John Rectenwald, Peter Henke, Thomas Wakefield, Lena Napolitano
No Product/Research Disclosure InformationUniversity of Michigan, Ann Arbor, MI